Chapter 9 Annual cycles in southern African weavers: breeding seasonality and moult patterns 182 Annual cycles in southern African weavers: breeding seasonality and moult patterns Introduction With 116 species, the weavers comprise a large family with a wide diversity of life histories, as highlighted by Crook (1964). Lack (1968) acknowledged that Crook’s approach to the study of ecological adaptations in weavers formed part of the inspiration for his classic text Ecological adaptations for breeding in birds. Of all passerine families, the weavers, along with the blackbirds Icteridae of North and South America, show the greatest diversity in breeding habits (Lack 1968). This diversity is also seen in the diversity of timing and duration of primary moult presented in this thesis. Moult has been studied in several southern African species previously, but only two papers (Oschadleus et al. 2000, Craig et al. 2001) used a rigorous statistical technique (the Underhill- Zucchini model) that allows precise comparisons between species and geographic areas. This chapter is an overview of the preceding chapters and summarizes the annual cycle of southern African weavers, in particular the timing of breeding and post-nuptial moult. First, however, the relative wing shapes of southern African weavers are discussed. Relative feather masses Weaver wing shapes are fairly uniform in that the primaries increase in size from the innermost and then decrease near the outermost primaries. There is variation on a finer scale as to which primary is the largest, and in the extent of reduction of the 10th primary. In the Ploceidae generally the degree to which the outermost (10th) primary is reduced is subject to much variation. Moreau (1960: 449–451) studied the length of the outer primary in a large number of weavers and found no correlation with taxonomy or habitat. The primary feathers of the wings were weighed, in order to calculate relative feather masses, for Sociable Weavers Philetairus socius, Chestnut Weavers Ploceus rubiginosus, Thick-billed Weavers Amblyospiza albifrons, Lesser Masked Weavers 183 Ploceus intermedius, White-winged Widows Euplectes albonotatus and Long-tailed Widows E. progne (Table 1). In addition, relative feather masses were obtained from the literature for Cape Weavers Ploceus capensis (Underhill and Joubert 1995), Southern Masked Weavers P. velatus (Oschadleus et al. 2000), Village Weavers P. cucullatus, Red-billed Queleas Quelea quelea and Southern Red Bishops E. orix (Craig et al. 2001). Primary moult parameters for Lesser Masked Weavers were not included in this thesis because the available sample size was small. Plotting the individual primary feather mass percentages showed an increase in percentage mass from Primary 1 to Primary 9, although the relative masses of Primaries 8 and 9 are similar (Figure 1). Primary 10 is small in all species, accounting for between 0 and 2% of total primary feather mass (Table 1). All weaver wings analysed showed a rounded wing, across different genera. Thick-billed Weavers, Long-tailed Widows, and to a lesser extent Sociable Weavers, however, had more rounded wings than those of other weavers, with the feather mass of Primary 9 being less than Primary 8 (Figure 1). These three species differ in size and are in different genera. Within the four Ploceus species there was little variation in relative feather masses (Table 1); most of the variation related to the extent of reduction in Primary 10. Craig et al. (2001) did not provide masses for Primary 10 in Village Weavers; in this species Primary 10 is vestigial (pers. obs.). Potential factors influencing wing shape are protection against physical abrasion and aerodynamics (take-off, migration, flight displays). Thus Sociable Weavers may have rounded wings to protect the outer primaries from abrasion on their nests (see Chapter 2). Dawson (2005) suggested that, in a range of European passerines, that the greater relative mass of the outer primaries in some species may reflect a protective role against physical abrasion, or an aerodynamic role in that each of these feathers provides a leading edge to the wing. He found that scaling relationships (log mass vs log length) were related to flight characteristics and habitat, rather than to phylogeny. In European Starling Sturnus vulgaris, take-off parameters vary with wingtip shape; birds with more rounded wingtips tended to take off from the ground at a steeper angle of ascent than those with relatively more pointed wingtips (Swaddle and Lockwood 2003). Given the wide variety of flight activities of the weaverbirds, an inter-species 184 analysis of wingtip shape with the ploceids is likely to be a rewarding avenue for future research. Wing shape is related to migration, with long-distance migrants having more pointed wings (Underhill and Joubert 1995). Weavers are not long distance migrants, and the longest movements occur in Red-billed Quelea (Jones 1989). Chestnut Weavers are resident in some areas but show regular movements in others, often correlated with rainfall (Fry and Keith 2004); the longest known movement is 213 km in East Africa (Backhurst 1977) and 284 km in southern Africa (Oschadleus and Brooks 2005). These two species, Chestnut Weaver (males) and Red-billed Quelea, have the most pointed wings of the weaver species for which data are available, with the relative mass of Primary 9 being greater than 14.5% (Table 1). Long-tailed Widows have larger wings in males than in females to compensate for the aerodynamic costs of a large tail in the male (Balmford et al. 1994). This may also be a reason for the more rounded shape of the wing in this species compared to other species. The difference in relative primary feather masses between males and females is worth investigation. The only species in which only the male has a long tail and for which feather mass data for both sexes are available is the Cape Sugarbird Promerops cafer (Underhill and Joubert 1995); their data shows considerable differences between the sexes. Male sugarbirds use their wings to produce a snapping sound in display flights, and this may explain part of the difference in relative primary masses between males and females in this species (Skead 1967). Underhill and Joubert (1995) modelled wing shapes using relative masses of primary feathers, using second order Tchebycheff polynomials in a regression analysis, to describe the shape of a bird’s wing with two parameters. The same method was employed here using the first nine primaries of the weavers. The advantage of using Tchebycheff polynomials is that the estimates of the regression coefficients are uncorrelated with each other (Underhill and Joubert 1995). This method produces three regression coefficients, denoted R, S and T. Underhill and Joubert (1995) showed that the coefficient R is of no interest in describing wing shape, and that S and T summarize the manner in which relative primary masses change. A plot of S vs T thus summarizes the nine-dimensional relative feather mass data in a two-dimensional plot; Underhill and Joubert (1995) plotted 185 S vs –T, and this convention is followed here. If a species has both S=0 and T=0, all nine primaries are of equal mass; if both S and T of a species are near to zero, so that are close to the origin in the plot, the species must have a relatively uniform set of primaries. Figure 2 provides a plot of S vs –T for the 11 species of weavers for which data are available. The Red-billed Quelea has a relatively large positive value for S and T is nearly zero; Underhill and Joubert (1995) show that this indicates that the feather masses form a geometric progression from the smallest (innermost) to the largest (outermost) primaries. Both male and female Chestnut Weavers have large positive values for S and small negative values for T (Figure 2). This indicates that the rate of increase in mass decreases for the outermost primaries; however in this species the outermost primary is still the heaviest. The Thick-billed Weaver and Long-tailed Widow have a relatively small positive values for S and large negative values for T; this is characteristic of species for which the outermost primaries are not the largest. Most of the remaining weavers in Figure 2 have intermediate values for both S and T, indicating that that the rate of increase in mass decreases for the outermost primaries, but several of the outermost feathers may be similar in mass. The potential of the approach to describe one aspect of wing shape pioneered by Underhill and Joubert (1995) should be further explored with a larger sample of species. Breeding in southern Africa weavers In the eastern parts of southern Africa, especially KwaZulu-Natal, peak summer rainfall is in December; farther north in Zimbabwe, the peak is in January; to the west, in Botswana and Namibia, peak rainfall occurs in late summer, in January and February (Allan et al. 1997; see Figure 1 of Chapter 1). The overall pattern is that the time of peak rainfall across the summer rainfall region of southern Africa moves in an anticlockwise direction. Most granivorous birds in the summer rainfall region breed in mid- to late summer, in response to rain which results in a flush in insect abundance to feed their young and also results in seed crops from rains several weeks earlier (Skinner 1995). The peak breeding was plotted as a median for each species per region, using the data from the BirdLife South Africa Nest Record Card Scheme (NRC) (RP Prs-Jones 186 and I Newton unpublished data). Peak breeding in the Western Cape was in September for the four weaver species that occur there (Appendix 1). Peak breeding for all other regions was more spread out for different species, and occurred mainly in summer with some records in late spring or autumn (Figure 3).